月球车地球敏感器图像地心位置提取算法

地球是月球上可视半径最大的天体，且在天球上运动范围较小，不存在升降现象.当月球车在月球对地面区域活动时，利用地球敏感器对地球成像可实现月球车长期自主天文导航.地球图像地心位置提取是利用地球敏感器进行天文导航的关键技术之一，直接决定了地球敏感器的观测精度.本文通过研究地球敏感器镜头投影模型分析地球成像规律，提出一种不受地相变化约束的地球敏感器图像地心位置提取算法，采用取半搜索法和循环搜索法两步实现地球真实边缘线的充分筛选，并基于此拟合地心位置.半物理仿真实验校验结果表明，本算法能针对不同地相图像有效提取地球中心，外符合平均精度约为9.78"~16.68"，在实验条件随机改变的情况下，地心位置外符合精度标准差互差最大不超过0.98". 

Geocentric Position Extraction Algorithm for Earth Sensor Image of Lunar Roverormalsize

The Earth is the largest visible celestial body on the moon, whose range of motion on the celestial sphere is small. Moreover, it has no phenomenon of rise and set. When lunar rover is exploring on the moon surface facing toward the Earth, the Earth sensor can be used to photograph the Earth, thus achieving long term autonomous celestial navigation. Geocentric position extraction algorithm for Earth image is one of the key techniques of celestial navigation using Earth sensor, which directly determines the precision of the sensor.In this paper, the Earth imaging law is analyzed by studying the projection model of Earth sensor lens. An algorithm for extracting the geocentric position of the Earth sensor image free from Earth phase change is presented. Two steps are proposed to realize full screening of the real edge of the Earth, including semi search method and cyclic search method, and then geocentric position can be fitted. The semi-physical simulation experiments show that the algorithm can effectively extract geocentric positions for different phase images and it earns the external average accuracy of 9.78"~16.68". Under the circumstance of random change of experimental conditions, the maximum difference of standard deviations of external accuracy exceeds no more than 0.98".